Bayesian Econometrics Textbook

Zahid Asghar has found for me a very nice textbook, which covers the same material that I have covered in the past few lectures. It is also at about the same level of mathematical difficulty. Chapter 8 covers Bayesian inference in the Beta-Binomial model, while Chapter 9 compares the conventional frequentist inference to Bayesian. This is the reverse order of what I did in class. ALSO, the chapters go into much more detail which I have omitted in the lectures.

It would be very useful for the students to go through these two chapters. Both chapters have exercises at the end. I would like to assign these exercises to the class as a whole. DIVIDE up the exercises among yourselves and produce solutions. Each student should do one exercise, so that the class as a whole creates a solution to all the exercises. You can cooperate and collaborate, but everyone should participate. YOU ONLY learn by solving problems, NOT by watching others solve them.

I have attached the two relevant chapters to the webpage with the associated lectures:

See: BLO4: Conventional Inference for Random Sampling for Bayesian Statistics Chapter 9 which compares conventional and Bayesian Inference.

See BL06: Bayesian Inference in Beta-Binomial Model  for Bayesian Statistics Chapter 8 which covers exactly this topic.

Probably you should read these chapters in the proper sequence (first 8 then 9) which is the opposite of the sequence followed in the course. READ and try to understand these chapters because they develop more details of the issues that I could only cover very briefly in the lectures. ONE ESSENTIAL skill required for doing research is the ability to read and understand material on your own. WITHOUT this skill, there is NO chance of your completing a research thesis at either M Phil or Ph.D. level. WITH this skill, the sky is the limit. 

Methodological Mistakes and Econometric Consequences

This is not directly related to our course, but very important lecture on the wrong foundations of conventional econometrics. This is based on a paper which provides more details and discusssion:

Methodological Mistakes and Econometric Consequences

BL03 Beta Binomial

The third lecture is on the simplest Bayesian model, which is the Beta-Binomial model. Here the data is Binomial, while the prior is a beta density.  For useful notes on this model, please see lecture notes given on the following link (please read these lecture notes to prepare for class).

An introduction to the Beta-Binomial model - School of ...

https://www.cs.cmu.edu/~10701/lecture/technote2_betabinomial.pdf

by D Navarro - ‎Cited by 3 - ‎Related articles

The beta-binomial model (along with the closely-related beta-Bernoulli model) is probably the simplest interesting Bayesian model. It is tractable ..... calculations.

At the moment, the Wikipedia entry on the Beta-Binomial will be too difficult to read, but after having done the lecture or two on the concepts, you should be able to follow the discussion there as well:

https://en.wikipedia.org/wiki/Beta-binomial_distribution

BE L02 URDU: Foundations of Probabilty

BL01 URDU Motivational First Lecture

New Course: Feb 2016-May 2016

Bayesian Econometrics was cancelled last year because of low enrollment. Now we are offering it again this term Those who go through this course will acquire world class skills in Bayesian econometrics, which are very useful in doing data analysis for research related to M Phil and Masters theses. These methods allow new ways to analyze and interpret data which are not available with standard econometric techniques.

The course will focus on interpretation and implementation, and not proofs and mathematics. It is designed to be user friendly, and teach student how to DO Bayesian econometrics on real world data sets in context of applications. In comparison with current available textbooks, this approach is very light on mathematics. Nonetheless, some mathematical work is required. We have been very careful to select only the minimum essentials required, and also ensured that the material can be grasped by students with average background available here. IN RETURN, we ask the student to make the efforts necessary to understand the course.

The course materials will be available on following website:
Bayesian Econometrics

Before starting the course, some preliminary basic materials must be mastered. These are covered in lectures which are described here.

PRELIMINARIES: This material should be studied and mastered BEFORE starting the course

Preliminary Introductory Lecture 0: Islamic Approaches to Knowledge: 

Lect 1: is about drawing histograms for data  == Lec 7 of Introduction to Statistics == Lec 2 of Applied Econometrics

Lec 2: Making Kernel Density representation of data -- an improvement over histograms == Lec 8 of Intro Stats

Lec 3: Basic Theory of Binomial Random Variables - De Moivre's Theorem == Builds upon Lec 13 of Intro Stats, and parts of Lec 08 Intro to App Econometrics

Please study this material before the start of the course.

Asad Zaman

DHSY -- Detailed Explanation

The DHSY paper on the modeling of the consumption function is extremely important. It is very condensed and contains a lot of information packed into it, which is not explained in detail. It is worth reading carefully, and can be the basis of a full course on econometrics. EVERY econometrician should know the paper thoroughly as it discusses most important basic concepts required for the knowledge of econometrics. Non-econometricians can also benefit from reading and understanding it.

At the moment, I am planning to go through paper with my students, where we can take turns in reading and interpreting. I have set up a website to which we can post, and also add comments at the bottom of the pages for discussion.

The website is: DHSY

There are two panels on the pages. One side should contain the original text, while the other side should contain the interpretation, to be written in by students in turn.

The original paper is in PDF format, which makes it difficult to copy from. You can get it into word format by using OCR -- Optical Character Recognition   First you have to copy a page from the pdf file and put it into Windows PAINT. Then save the file as JPG format and then upload to this website for character recognition. Probably wont work with formulas and graphs and so those should be typed it or else cut and pasted as graphics.